Search

Abstract The online classification of grid disturbances is an important prerequisite for an automated and reliable operation of power transmission systems. Most of the state‐of‐the‐art approaches assume that all classes are already known in the training phase and cannot handle new disturbance events, which appear in the application phase and lead to severe misclassifications. To mitigate this shortcoming, the disturbance detection is investigated as an open classification task and a novel recurrent Siamese neural network architecture is introduced to identify and locate known and unknown disturbance events from phasor measurements. Extending preliminary work, a probabilistic distance‐based classification approach with an integrated rejection mechanism is presented, which enables to learn class‐dependent decision boundaries and margins to reduce the open‐set risk. A detailed performance analysis is presented including multiple benchmark methods in different closed‐set and open‐set classification tasks for a simulated power transmission system. Additionally, a limited and full observability of the grid with phasor measurements are addressed in the experiments.

Abstract The automated identification and localisation of grid disturbances is a major research area and key technology for the monitoring and control of future power systems. Current recognition systems rely on sufficient training data and are very error‐prone to disturbance events, which are unseen during training. This study introduces a robust Siamese recurrent neural network using attention‐based embedding functions to simultaneously identify and locate disturbances from synchrophasor data. Additionally, a novel double‐sigmoid classifier is introduced for reliable differentiation between known and unknown disturbance types and locations. Different models are evaluated within an open‐set classification problem for a generic power transmission system considering different unknown disturbance events. A detailed analysis of the results is provided and classification results are compared with a state‐of‐the‐art open‐set classifier.